Alzheimer's disease (AD) is characterized by profound impairment in memory and other cognitive skills that make it difficult for patients to complete instrumental activities of daily lving (IADL) and live independently. Mild Cognitive Impairment (MCI) describes a condition in which an older adult shows poor memory, but is generally able to perform IADLs. Executive functioning (EF) refers to the ability to plan, reason, solve problems, and multi-task, and is dependent upon fronto-parietal networks. Functional imaging studies have documented increases in prefrontal cortex activity in MCI, and paradoxical increases in the connectivity between memory networks and fronto-parietal regions. We hypothesize that MCI is characterized by reorganization of frontally-mediated networks to compensate for AD pathology. Using EF as a means of defining fronto-parietal networks, the work proposed here investigates the structural and functional connectivity of prefrontal networks in MCI and AD. In aim 1, we will investigate if increases in brain activity in fronto-parietal networks reflect impaired structural connectivity within the network. In aim 2 we will explore how increases and decreases to the functional connectivity of frontally based networks predict IADLs in MCI and AD, respectively. We will enroll 55 patients with MCI, 55 patients with AD, and 55 age matched elderly controls (EC). All participants will complete a comprehensive clinical assessment and undergo MRI scanning on a 3T magnet. Our first specific aim is to test the hypothesis that in MCI, declines to the structural connectivity of fronto- parietal networks will correlate with increases in brain activity in this network. This would support the theory that declines in network structural connectivity induce changes in functional activity to preserve cognition. Toward this aim, MCI and EC will complete cognitive fMRI tasks of EF and diffusion tensor imaging (DTI). We will use DTI tractography to create the white matter tracts that connect the fronto-parietal networks involved in the tasks. We will extract mean fractional anisotropy values (FA) from these tracts to interrogate their integrity. We will correlate FA with fMRI brain activity to understand how changes to structural connectivity impact the network's functioning. Our second aim will test the hypothesis that there will be increased functional connectivity in frontally mediated networks in MCI, but reduced functional connectivity in AD. Moreover, in MCI, increased functional connectivity will correlate with poorer EF and IADLs. This finding would support the theory that in MCI, increases in functional connectivity reflect reorganization of frontal networks as a compensatory mechanism. In order to test these hypotheses, we will first identify frontal-parietal networks engaged during EF fMRI tasks in MCI and EC. Next, MCI, EC, and AD will complete resting state fMRI (rsfMRI). We will compare the functional connectivity of frontally based networks between groups and explore how functional connectivity is associated with clinical symptoms. The results of both aims would support the theory that MCI is a dynamic state characterized by increases and decreases to the connectivity of neural networks. The seemingly paradoxical increases in brain activity and functional connectivity of frontal-based networks may be an attempt to compensate for accumulating AD pathology in memory networks. AD diagnosis and impaired IADLs may occur when the structural connectivity of fronto-parietal networks can no longer support compensatory functional activity (aim 1), and functional connections can no longer offset impairing symptoms (aim 2). This work will provide new insight into brain re-organization in MCI. It will support the integration of a nonlinear trajectory of brain changes int neuroimaging models of AD disease progression, lead to the development of disease resilience biomarkers, and begin to characterize new targets for treatment focused on supporting compensation.